Reproducibility of Radiographic Measurements Made in the Active Stages of Legg-Calvé-Perthes Disease: Evaluation of a Prognostic Indicator and an Interim Outcome Measure

Abstract

Background: Deformity index (DI) and extrusion index (EI) are 2 radiographic methods currently used to quantify femoral head deformity in the active stages of Legg-Calvé-Perthes disease. This study aimed to quantify the interobserver reproducibility of the 2 methods using a large, international group of pediatric orthopaedic surgeons with diverse clinical experience.

Methods: Radiographs of patients (age 6 to 11 y at time of diagnosis) prospectively enrolled in an international multicenter-cohort study, were measured by members of our study group. For each radiograph, the raters independently assessed EI (n=33 cases) and DI (n=32 cases), along with the rater's subjective assessment of the extent of hip deformity (ie, none, mild, moderate, or severe). Reliability analysis was conducted using intraclass correlation coefficient (ICC) and κ techniques.

Results: The ICC for EI on the affected side was 0.68 (95% confidence interval: 0.57-0.79). The calculated DI ICC was 0.53 (95% confidence interval: 0.41-0.68). Subjectively, an average of 68±3.5% of surgeons agreed on the subjective description of each patient's femoral head deformity.

Conclusions: EI is measured with a reasonably high rate of correlation among surgeons from disparate backgrounds. Agreement is lower among the same group of surgeons when the more complex DI is used. DI is most useful when utilized by a small number of experienced observers in a research setting, whereas EI may have better clinical applicability.

Level of evidence: Level III-reliability study.